Can custodians reinvent themselves as networked data businesses?
What is the future operating model of the custody industry? A presentation by Dominic Hobson at the Global Custody Forum in London on 2 December 2014.
I chose to play it this morning not just because I have not listened to any popular music since I was 21.
Nor because the words seem eerily relevant to an audience of bankers.
“You are wondering now, what to do, now you know this is the end. You are wondering how you will pay, for the way you did behave.”
As somebody said to me last night, it would be appropriate to play it at the funeral of a banker.
And we may be about to attend exactly that – metaphorically speaking.
For the story of The Specials provides a cautionary tale for Custodians.
A couple of years ago I ran into the remnants of The Specials in the departure lounge at Porto airport in Portugal.
As usual, the inbound Easy Jet flight was delayed, so we had plenty of time to talk.
They had just performed at a music festival in northern Portugal.
Because their earnings from recorded music had dried up completely.
To get paid, they had to play live.
In fact, they had reformed the band in 2009 for precisely that reason.
What had happened to them, of course, was Napster – and its successors.
Now Napster was not even invented until 1999.
But within less than 10 years digital file sharing of recorded music was powerful enough to reduce pop stars to flying by Easy Jet.
Now journalists, as you know, have always flown by Easy Jet.
That apart, technology has done exactly the same thing to journalism as it did to recorded music.
Many people complain to me about the declining quality of the editorial in specialist financial journalism.
There is a reason for that.
It is that the revenue which paid for high quality journalism – namely, advertising – has evaporated.
Publishing has been digitised.
And the question I would like to explore today is this: why should not custodian banking be digitised too?
On this slide I have put a random list of what custodians know
Every single item on this list can be reduced to some combination of digits.
It is worth asking why custodians are getting paid for processing those combinations of digits.
After all, the marginal cost of processing data in digital form is effectively zero.
Yet here are just three global custodian banks being paid, between them, $12½ billion a year to process digital data.
Just two of them are paying dozen of sub-custodian banks another $1 billion a year to do the same thing at the local market level.
The global custody industry is probably making $25 billion a year for processing data whose marginal cost is zero.
Why then, you might ask, is the industry not more profitable?
One reason is that custodians do not have their costs under control.
As this chart shows, the bill for staff salaries and benefits is rising, not falling. 
This is partly because compliance departments are expanding.
But is mainly because people work in banks for money rather than pleasure, and they expect to be rewarded handsomely when revenues rise.
As you can see, average annual salaries and benefits have, basically, never stopped rising.
But it is also because custodians are not using technology to raise productivity.
They use technology to cut prices, not to cut costs.
The best way of measuring this is to note that, despite ostentatious staff reduction programmes, the number of staff has actually gone up.
If you use technology to cut prices rather than costs, taking on more assets inevitably means hiring more people.
And more people mean higher costs.
It is those higher costs that are rendering the custody industry vulnerable to a disruptive technology.
A disruptive technology is a technology that poses an existential threat to an industry by totally transforming its commercial economics.
The company failed to make the transition to the digital age not because it had failed to understand the technology.
Its researchers had actually invented digital photography.
It failed to adapt to the new environment created by digital technology because it was too big and too blundering and too political.
The wrong people were in charge.
The wrong people were rewarded.
The wrong people were promoted.
I am not saying your bank is like that.
I am saying I would not be surprised if it was.
At its peak, Eastman Kodak was worth $28 billion and employed 140,000 people.
When Instagram was floated for $1 billion in 2012, it employed just 13 people.
That means that each Instagram employee was 385 times more productive than each Kodak employee.
That is the power of digital network technology.
But instead of responding to the gravity of this threat, the global custody industry is making no serious attempt at all to adapt its business model to the digital age.
On the contrary, I have detected more than a little self-congratulation this year as fee income has risen.
As you can see, there has been an uptick in servicing fees – particularly lately.
And in sub-custody fees too – for some at least.
This rising income is an illusion.
Something would be seriously wrong if banks paid largely through ad valorem fees were not seeing an increase in revenues when stock market values increased.
In fact, as this chart shows, there is a pretty strong correlation between rising stock markets and rising servicing fee income
Unfortunately, stock markets have risen not because earnings and prospects of earnings in the real economy have improved.
But because money is cheap and it is plentiful.
As this chart shows, central banks have since the crisis created a deluge of digital money, and used it to buy securities.
So of course the prices of securities have gone up.
What custodians are enjoying is not an improvement in business conditions but inflation in the price of financial assets.
As Quantitative Easing is withdrawn, we can expect stock market prices to come back into alignment with the real earnings of the corporate sector.
But that is not happening yet.
One useful measure of the extent to which central bank money is still propping up asset prices is the shrinkage of the tri-party markets.
Prior to the financial crisis, the tri-party market was the principal venue where investment banks raised finance from cash-rich banks not just to fund their own assets, but to fund the assets of their buy-side clients.
As this chart shows, the tri-party market in the United States peaked in the Spring of 2008 – about the time Bear Stearns was bailed out by J.P. Morgan.
It has pretty much flat-lined for the last five years.
The same is even truer of the European tri-party repo market.
The relatively slow growth of the Global Securities Financing businesses at Clearstream and Euroclear ovder the last five years can be seen in this chart.
The apparent surge at Euroclear in the last two years is something of an illusion, in that it largely reflects the opening of its “Collateral Highway” to assets in custody other than at Euroclear.
This reflects, of course, central bank activity.
In Europe, the Long Term Re-financing Operation (LTRO) by which the European Central Bank has financed banks in the eurozone since 2011 has displaced commercial bank funding via the tri-party market.
LTROs in Europe are still running at over €500 billion on the ECB balance sheet.
In the United Kingdom and the United States, the central banks have since the acute phase of the crisis passed preferred printing money to buy securities to funding banks directly – so investors now sell their securities to the central banks, not the commercial banks.
Nothing illustrates this better than this chart.
It shows that tri-party in the United States has shrunk as central bank buying of securities has risen.
The gap between the two lines is a measure of how central bank money has displaced the commercial bank money of the much reviled “shadow banking” industry.
In other words, all the extraordinary monetary policies of the last five years have done is: -
Transfer the holding of government debt from investors to banks.
And replaced commercial bank funding with central bank funding.
The interesting question is: will the commercial bank money come back as central bank money is withdrawn?
The effects of this policy are plainly visible on the balance sheets of the custodian banks.
Look at how cash has piled up on the balance sheets of BNY Mellon and State Street since the financial crisis.
As this slide shows, cash deposits at BNY Mellon and State Street have more than doubled since 2008.
What banks look to do – even custodian banks – is make more money on their assets (their loans) than their liabilities (their deposits).
Unfortunately, record low interest rates – combined with regulatory pressure to sacrifice yield for quality and liquidity - have compressed that differential.
As this chart shows, net interest margin has been declining for years.
So we have the paradox of custodians with lots of cash on their balance sheets.
While net interest margin flat-lines.
This problem is going to get worse before it gets better.
The reason why is the Liquidity Coverage Ratio (LCR).
Like any bank, custodians are re-positioning the asset side of their balance sheets to take account of the LCR, which starts to come into effect one month from today.
The LCR forces banks to hold assets liquid enough to pay off their most unstable sources of funding without difficulty for 30 days.
Inevitably, that means more deposits with banks and central banks, and more government or government-guaranteed securities.
These are showing up already on the asset side of the balance sheet.
Most of those securities holdings are government or government-guaranteed bonds, which attract a lower capital haircut.
Just as cash deposits make up two thirds to three quarters of custodian bank liabilities so cash deposits and government securities make up two thirds to three quarters of custodian bank assets.
And those numbers are large.
$90 to $130 billion in deposits at each of BNY Mellon and State Street.
$115 billion in government securities at each bank.
As tapering starts to bite, and interest rates to rise, the prices of those securities is going to fall – forcing the custodians either to reduce short term liabilities or increase short term assets.
Neither will help them make money.
It gets worse.
A little further off than January 2015 lie two further taxes on Net Interest Margin.
One is higher haircuts on “non-operational” deposits.
That means a lot of those hot money inflows seeking a safe haven are not going to be worth having any more.
The second is the Net Stable Funding Ratio (NSFR), which will come into effect on 1 January 2018.
The NSFR does for long term assets what the LCR does for short term assets: bring the duration of liabilities into line with the duration of assets.
What all this adds up to is:
More expensive funding.
Less remunerative assets.
And lower Net Interest Margin.
The consequent scramble for higher quality assets will drive their prices up.
With negative knock-on effects on the market for general collateral in the securities lending and financing markets as well.
These effects are much more visible already at one group of banks which would be horrified to be considered custodians, though they are.
I refer to the prime brokers.
The immense profitability of prime brokerage in its golden age depended on its ability to re-use without let or hindrance the cash and securities prime brokers held in custody.
They borrowed the cash to fund their own businesses.
They lent assets owned by one client to another.
They pledged client assets as collateral in the repo market.
And then lent the cash back to the same clients.
If you have to match the duration of assets to liabilities, that trade no longer works.
As it happens, the trade has not worked for the stand-alone prime brokers for years.
As this chart shows, the prime brokerage revenues of Goldman Sachs have never recovered from the crisis – chiefly because Goldman Sachs can no longer borrow money overnight and lend it to a hedge fund for 90 days.
It helps to explain why all of the universal banks - BNP Paribas, Citi, Deutsche, HSBC, J.P. Morgan – have, in their different ways, chosen to merge their custody division with their prime services.
Some of you will already have experienced at first hand what happens when the investment bankers take over.
But the theory is impeccable, even if the practice is unpleasant.
In theory, assets can be held in a single account, from which they can be serviced, financed - and indeed lent.
The only problem is that less and less people want to borrow them.
As this chart shows, the value of securities on loan with two major agent lenders has been flat for the last five years.
There are so many forces arrayed against a pick-up in stock borrowing that it is hard to see how it can ever recover.
Extraordinary monetary policies have distorted stock loan markets too.
Low interest rates have shrunk returns from cash collateral reinvestment.
Corporations are so cash-rich they buy each other with cash, eliminating trades that try to profit from trading the relative prices of the cash and stock offers.
Then there is regulation.
Permanent restrictions, such as those imposed by the European Short Selling Regulation.
Single counterparty limits in Dodd Frank.
Additional weighting of large exposures in Basel III.
Both of those will hit custodian indemnities to institutional lenders directly.
The Volcker Rule: by eliminating prop trading, it has eviscerated demand to borrow general collateral.
Even the classic dividend trade is riskier, and less profitable, as the authorities attack tax avoidance.
This is the result: revenues running at a fifth to two fifths of pre-crisis levels.
The sensible choice for many lenders is to stop lending.
And many have.
To maintain revenues at all, custodians are taking more risk: they are doing more principal business. 
Something similar is observable in the other great trading business of the global custodian banks: foreign exchange.
This chart records a recent uptick in FX revenues.
That, too, reflects an increasing appetite for principal business.
Increased willingness to commit capital, and to speculate.
With banks being fined for manipulating FX rates, it is hard to think of a less sensible time to engage in currency speculation.
Recent developments in transition management imply that some custodians have yet to change their culture sufficiently to be confident of escaping a regulatory accident.
In fact, increased principal activity is itself a direct consequence of client concerns raised not just by regulatory investigations, but by a series of law suits launched since 2009 by high profile clients who alleged that their custodians had exploited their currency trades.
It is precisely because FX rates are now being disclosed, and can be managed downwards by well-informed clients, that custodians are having to speculate more.
What, you might ask, do all these observations amount to?
In cash management, it means the regulators now choose the depositors.
They are also the primary influence over net interest margin.
In securities lending, it means there are fewer and fewer opportunities to lend.
Indemnities are unaffordable.
Fee splits will never get more generous to the banks.
Either risk must be increased massively, or vast investments made in complete re-automation of the securities lending process from top to bottom.
That is impossible to do without withdrawing from the market for a year or two.
So it cannot in practice be done.
In foreign exchange, the clients no longer trust the banks.
Thanks to the litigation, and the scandals, more and more are executing FX elsewhere.
Those that have stayed are paying less.
And, what is in some ways worse, demanding proof that they are paying less.
The only solution is risky and capital-intensive speculation as a principal trader.
But the really bad news is that the core business of this industry – the processing of securities transactions, and the servicing of financial assets – is also going away.
It is moving, slowly but inexorably, from banks to utilities.
Custodians have feared this for as long as I have been writing about this industry.
The difference is that now it is really happening, and it is unstoppable.
Clearing is already the preserve of the CCPs.
Settlement the preserve of the CSDs.
As CSDs become banks – and banks become CSDs - sub-custody is heading for extinction.
The settlement of funds is already controlled by CSDs in France and the United States.
The rest of the world will follow suit, eliminating the transfer agency business.
Collateral management is already largely controlled by the ICSDs.
And what remains will be captured by them and the CCPs, as OTC derivative clearing begins in earnest in 2015-16.
The securities lending and repo markets, now being suppressed as a deliberate act of policy, will survive only as the preserve of infrastructure.
CCPs already clear both.
CCPs are also in the process now of devouring the clearing fees that were once the preserve of the clearing brokers in derivatives markets.
The first KYC, AML and sanctions screening utilities are now being built, as infrastructures.
They will encompass the final extinction of the custodian-friendly omnibus account, which is dying, where it is not dead already.
We can expect the last great source of inefficiency in post-trade services – namely, corporate actions– to be eliminated by a combination of compulsory standardisation and data utilities.
In short, the only future in prospect for the global custodians is to act as the risk takers, of first and last resort, to infrastructures.
If you do not believe that, examine the duties of a depository bank under AIFMD and UCITS V.
Ask why collateralised exposures to infrastructures win capital relief.
And ponder why regulators are now talking not about the recovery and resolution of banks, but about the recovery and resolution of infrastructures.
For the global custodians, Othello’s Occupation is going, where it is not already gone.
It is not a particularly compelling proposition, which helps explain why the stand-alone custodians have underperformed the broader markets.
The interesting question is what custodians should do about it.
One answer, which at least one stand-alone custodian has been bold enough to adopt, is to become an infrastructural utility yourself.
Others are betting on a more daring response.
This is to become a data business.
It is an understandable ambition for a bank - and not just a bank facing the extinction of its traditional business either.
Anyone looking at the astonishing and rapid success of the data-driven network businesses is bound not only to be impressed by their success, but to be more than a little afraid of it.
As this chart shows, Facebook, Apple, Google, Amazon and PayPal make the customer bases of even the mightiest banks look comically small.
If any one of these networks chose to enter any of the businesses currently conducted by the global custodian banks – safekeeping, asset servicing, fund accounting, transfer agency – they would start with formidable advantages.
And especially of reputation.
They also know custodians possess large volumes of the substance that drives their businesses: data.
They know that data may well contain information or patterns or hidden or unexpected correlations that can be turned to profit.
Or which might illuminate the path to the development of new and lucrative products and services.
So, for the custodian banks, testing these hypotheses themselves has excellent defensive qualities.
It counters a competitive threat.
And it provides a good reason to get closer to clients.
After all, new products and services are unlikely to spring fully formed from the minds of custodial strategists.
New products are much more likely to emerge from day-to-day problem-solving on behalf of existing clients.
The interesting question is: what might those new products be?
One is aggregation of data: and its conversion into reports to investors and regulators.
A second is data mining: the search for correlations that can inform and improve decisions.
A third is to use the data to create new investment vehicles, such as index derivatives.
A fourth is new risk management tools, based on faster and more complete sets of data.
A fifth is assessing counterparties, such as funds or fund distributors, for their contribution to profits.
A sixth – and in many ways the most interesting – is to transform how clients trade cash, foreign exchange, equities, fixed income, futures and options, and OTC derivatives.
If that sounds counter-intuitive in a post-trade business, think again.
The distinction between the front office and the back is a linear idea from the analogue age.
In the digital age, it disappears.
In the digital age, data becomes part of a complex, adaptive system.
It becomes a circular stream of information which flows back into itself, continuously, creating constant negative feedback.
Prices and influences on prices feed asset allocation and risk management algorithms which inform decisions which need to be executed, which create prices and influences on prices which feed asset allocation and risk management algorithms, and so on, in an endless, circular stream.
It sounds compelling.
Yet reinventing a global custodian bank as a data-driven network business faces some formidable obstacles.
For a start, the customers will ask who owns the data.
If they think they do, they will want to be paid for it.
We see this happening already, in the resentment of fund managers at paying for data feeds from stock exchanges, Thomson Reuters and Bloomberg.
Once people want to be paid for data, a bank has to be very confident indeed that it can add value to it.
Doubtless regulators will also fret about the conflicts of interest associated with using client data.
Secondly, custodians will have to start thinking about what they do in a totally different way.
Up to this point, custodians have seen themselves not as custodians of data, but as custodians of securities.
Data does not get mined – or “curated” as the term now has it.
It gets processed.
Above all, data is not something that banks are used to sharing.
You do not have to work in the City or on Wall Street for long to learn that the most valuable commodity in any financial market is information.
In fact, the reason why finance is so profitable is that the information asymmetries are extreme.
Sellers always know more than buyers – and banks like to keep it that way, even with their favourite clients.
Effectively, bankers have used information to keep the least risky options for themselves, and to sell the most risky options to others.
All of the legal and regulatory problems we have encountered in this industry in recent years- in cash, FX, stock loan and, most recently, transition management – are, at bottom, the working out of the Information asymmetries that are the key to the historic profitability of this industry.
Unfortunately, or fortunately, depending on your point of view, in a properly networked, data-rich industry, those valuable information asymmetries are going to disappear.
Algorithms will visit and re-visit the data on which every decision is based and re-set the price, in real-time.
Buyers can no longer be told “Take it or leave it, this is what I do, and this is what you pay for it.”
Instead, sellers will have to adapt their products and services and price to multiple demands, and offer multiple ways of charging and paying.
You can see this happening already in one of the more technologically advanced parts of the industry.
I happened to look recently at the fee schedules of a group of fund platforms in Europe.
This chart shows how many different ways a group of just six European fund platforms are charging their users.
There is more than 300 platforms in Europe now.
And it will be data on costs and prices which governs the choice between them.
That is a revolutionary concept for a global custody industry which has until now used technology and data not to help clients make better informed choices, but to keep them captive.
Ask yourself, for example, why so few fund managers who have outsourced their back and middle office to a global custodian have managed subsequently to change their provider – and why it never takes them less than two years to complete.
When the Financial Conduct Authority here in London asked global custodians and their buy-side outsourcing clients two years ago to explain how quickly they could change their provider if it failed, they explained that it was impossible.
Frankly, that answer is far too complacent to survive a properly networked, data-rich marketplace.
The real answer – standardised data, captured on third party servers – will eventually be given.
It is up to the global custodians to decide if that answer is given by them, or by somebody else.
Once data is properly captured, outsourcing service providers are going to be interchangeable.
Information systems will capture all of the relevant data in real-time, creating genuine mobility for customers.
Far from taking two years to change a fund accountant, it may take no more than two minutes or even two seconds.
In fact, there is no reason why NAV calculations should not become the subject of a continuous intra-day auction by fund managers, in which fund accounting service providers bid for the right for their algorithm to be used to calculate NAVs.
The prices they charge will probably fluctuate in line with the availability of independent sources of processing capacity.
Investors will buy funds at real-time NAVs.
They will settle them in real-time too, and gross.
The current uncertainty over how many units in a fund a subscription actually buys will disappear.
In a future like that, there is no need for fund accountants as we know them today: just machines to perform calculations on data – and they do not have to be owned by banks.
There is no need for fund managers to rack up additional costs hedging portfolios against delays in receiving subscription monies or paying redemptions.
And there is no need for transfer agents at all.
Come to think of it, there is no reason why local settlement services should not be bought in exactly the same way.
By continuous, real-time auctions, in which sub-custodians compete for business with algorithmically powered calculation engines.
For sub-custodians, their future might be as bleak as that of the food manufacturers that supply the supermarkets.
They are now actually paying the supermarkets to distribute their products.
There is no reason of principle why what we now call sub-custody should not end up doing the same thing.
Far from escaping extinction through using powerful technology to “curate” data, a truly data-rich, networked securities services industry may have no need of custodians at all.
Which brings me back to where I began: with The Specials.
“You are wondering now, what to do, now you know this is the end.”
My own advice is this.
Stop thinking about the data.
Start thinking about the people whose activities actually create the data.
Not in the sense of “Thank goodness there are still people stupid enough to take the other side of my trades.”
Though there are plenty of them, especially – as I said last year – in the pension fund industry.
No - think of them instead as people with who you can build a relationship.
Whether custodians like it or not, the financial markets are moving towards a digital future in which their informational advantages have disappeared.
It is no longer a world in which – to quote from a Final Notice issued by the FCA – transition management executives will find themselves exclaiming to each other, “Nice!” and “Back up the truck!”
Opacity is out.
Disclosure is in.
And trust is essential.
In an industry such as banking, where trust has broken down is so many areas in so many different ways, the radical options are the safest options.
Have a look at this last slide.
There are many reasons why some banks are trading at discounts to book value, not least the reluctance of large European banks to face up to the quality of certain of their assets.
But there is clearly a deeper scepticism about the ability of large banking conglomerates to deliver a return on equity that exceeds their long term cost of capital.
A clear message from this chart is that the degree of scepticism is lower when a bank is smaller, more specialised and less conflicted.
In fact, those premiums commanded by BNY Mellon, Northern Trust and State Street would probably be higher still if they ditched their investment management arms.
But no CEO is going to take my advice on that, so let me close by making some more practical suggestions.
First, be prepared to pay for data, as well as to get paid for processing it.
The banks that pay clients to contribute data will acquire more clients, and so more data.
They will also attract a higher degree of trust – and even put themselves on the moral high ground relative to Facebook and Twitter and Amazon and Apple, which have never paid anyone for any of the data they use.
Secondly, as I recommended last year, banks should not resist increased equity capitalisation, but embrace it, evangelically.
Transform your balance sheets from 7 per cent Tier One equity to 47 per cent or even 77 per cent.
Of course, it will take time to accomplish, but it will create the kind of reassuring solidity and the reliable returns which make an institution not just creditworthy, but trustworthy.
Thirdly, never be a principal: always be an agent.
Don’t try and go back to a world in which banks bought all the sells and sold all the buys, and set the price at the one most advantageous to themselves.
Aggregate and net deposits, and currency pairs and stock borrowings.
Then display to the client a fully transparent set of live, executable prices.
That puts the client in control.
It also abolishes forever, and far more completely than any amount of law and regulation, the threat of litigation and regulatory fines.
In short, it wins trust.
And for the trustworthy agent, custody of the assets of the client is no longer necessary.
The whole market, for money, for currencies and for stocks, will be open to it.
In time, trustworthy agent banks will make far more money than their predecessors ever made surreptitiously over-charging captive custody clients.
Fourthly, get rid of cross-subsidisation.
Charge a proper price for the services you provide.
And be open about how you arrived at it: share full details of what goes into the price with clients: fees, commissions, spreads, all of it.
There is no point resisting the drive to disclose.
Digitised financial markets make possible a much greater degree of disclosure – and the smarter clients know it.
The successful custodian bank of the future is the one that is trusted.
And it is trusted because it is open; it is networked; it is unconflicted; it is collaborative; and it is extremely well-capitalised.
The generation of custodians that thinks and acts like this might even get away with a lot less regulation.
The quantity of regulation, like the quantity of law, is a measure of the degree of self-control in an industry.
Even regulators know that commercial agreements are more effective and enforceable at lower cost than rights granted by law.
And that accurate accounting and full disclosure is preferable to detailed regulatory prescription.
Just as good money can drive out bad, so good banking can drive out bad banking.
We will not be around to see it, of course.
But I like to think that this generation can at least begin the process of making custodian banking once more a fully human activity: dignified, truthful and free.
 $124 billion in custody at 20 per cent.
 Of course, both BNY Mellon and State Street have made major acquisitions during the period shown.
 The extraordinary collateralised funding measures introduced by the Federal Reserve at the height of the financial crisis included the Term Auction Facility of December 2007 to April 2010 (whose value peaked at $493 billion); dollar liquidity swaps with foreign central banks from December 2007 to February 2010 ($586 billion); the Term Securities Lending Facility of March 2008 to February 2010 ($236 billion); the Primary Dealer Credit Facility of March 2008 to February 2010 ($130 billion); the Asset Backed Commercial Paper Money Market Mutual Fund Liquidity Facility of September 2008 to February 2010 ($152 billion); the Commercial Paper Funding Facility of October 2008 to April 2010 ($350 billion); the Term Asset Backed Loans Facility of November 2008 to June 2010 ($48 billion). A Money Market Investor Funding Facility ran from November 2008 to October 2009 but was never used. That is a total of $1.995 trillion. The TALF (which can still be found on Federal Reserve publication H.4.1) was not finally repaid until the first week of November 2014, but all sums advanced under these measures are now fully repaid. These extraordinary measures should not be confused with normal discount window operations, or what the Federal Reserve calls Primary, Secondary and Seasonal Credit. The total value of the collateralised loans advanced via these three means, and outstanding to over 1,000 American banks at end-September 2012, was $4,064,244,000 (the data per bank is only published with a two year time lag). Currently (Wednesday 19 November 2014) H.4.1 records $248 million in such loans.
 The LCR is being phased in from 1 January 2015, with full implementation from 1 January 2017.
 The LCR introduces an as yet unresolved distinction between “operational” and “non-operational” deposits. Until now, custodians have not had to ask why a bank is depositing money with them. Under LCR, they do. The exact definition remains unclear but it is reasonable to suppose that operating deposits will be those that are a by-product of the core service of the bank: temporary accumulations of surplus cash thrown off by securities transactions which the custodian reinvests in time deposits, repos, commercial paper and money market funds, which may or may not be managed by the custodian, in conformity with guidelines agreed by the client. It adds incremental revenue to a fund, but managing cash of this kind is not a stand-alone business for a custodian. These “operational” deposits are judged by Basel III to be much “stickier” and so a stable source of funding because they are tied in by the wider relationship. “Hot” money inflows from financial institutions seeking a temporary safe haven before being reinvested will probably be deemed “non-operational” deposits of exactly the kind that inflates the denominator of the LCR in a punishing manner. Any deposit judged “non-operational” is therefore by definition less attractive to a bank – because the less stable the deposit, the more expensive liquid assets it has to buy to maintain its LCR, and the worse its Net Interest Margin.
 The NSFR is likely to persuade custodians to take less cash deposits from their banking and fund management clients, and to buy more high quality liquid assets, such as government bonds and central banks deposits, whose prices will rise and yields fall as a result. That means continuing compression of Net Interest Margin.
 Effective from 1 November 2012. No uncovered short selling (“naked shorts”). Disclosure of net short positions of equity in any one company (0.5 per cent) and short sales of more than 0.2 per cent of a company, and every further 0.1 per cent above that. Disclosure of short sales of sovereign debt above 0.1 per cent of issues up to €500 billion, and above 0.5 per cent of issues larger than €500 billion.
 Section 165 of the Dodd Frank Act sets single counterparty exposure limits of no more than 25 per cent of capital for bank holding companies with more than $50 billion in assets (including repo and reverse repo and stock loan). With more than $500 billion in assets, the limit is 10 per cent to other banks with at least $500 billion in assets.
 This raises the capital costs of the price of being in the business at all: the indemnities given to clients who lend. Banks could try restricting indemnification to the most profitable customers. That will be those who are least capital-intensive. It is hard to do when clients can be important for reasons other than securities lending, especially when the same counterparty limits and make it commercially worthwhile to spread business across a wider range of counterparties. Some clients will probably agree to lend on an unindemnified basis. But they are almost certainly be exactly the clients few custodians want. You could try and adjust the fee splits in your favour. But the benefit of doing that will in almost every case be vastly outweighed by the capital costs of doing the business.
 State Street admits in its latest 10-Q that the uptick in securities lending revenue in the first half of 2014 was “mainly reflective of growth in our enhanced custody business, where we participate in securities finance transactions as a principal.” State Street also says in its 10Q that: “In our role as agent, the indemnified repurchase agreements and the related collateral held by us are not recorded in our consolidated statement of condition … In certain cases, we participate in securities finance transactions as a principal. As a principal, we borrow securities from the lending client and then lend such securities to the subsequent borrower, either a State Street client or a broker/dealer. Collateral provided and received in connection with such transactions is recorded in other assets and accrued expenses and other liabilities, respectively, in our consolidated statement of condition. As of September 30, 2014 and December 31, 2013, we had approximately $15.78 billion and $11.29 billion, respectively, of collateral provided and approximately $7.01 billion and $6.62 billion, respectively, of collateral received from clients in connection with our participation in principal securities finance transactions.
 In January 2014, State Street was fined £22.9 million by the FCA for transition management “failings.”
 The majority of foreign exchange trades undertaken for custody and non-custody clients in transactions are ones where the custodian acts as principal, and not as an agent or broker. As a principal, the bank earns a profit based on its ability to risk manage the aggregate foreign currency positions that it buys and sells on a daily basis. In theory, those risks are not great. Few trades are large enough to require substantial capital commitment. Traders have an excellent idea of the riskless price at which a deal can be done, and quote accordingly. In volatile markets, they simply widen spreads. But the declining earnings might just tempt the custodians to take positions in pursuit of speculative profits. Regulators are investigating the FX markets, like the money markets, for alleged manipulation. Custodian entered the FX litigation crisis early, fielding the first of a spate of law suits alleging exploitation as long ago as 2009. That litigation long ago alerted clients to the spreads exacted by custodians. Most clients are now executing foreign exchange elsewhere or in ways less profitable to banks, such as standing agreements on the spread. To maintain foreign exchange earnings, custodians are taking on more risk. And there are lots of ways to lose money in the FX market. Such as short-selling currencies expected to decline in value. Betting that weight of money will move currencies in a predictable direction. Betting that high yielding currencies will pay more in interest than they lose in value. Carry trades vary, but al are ultimately bets that a higher yielding currency will not depreciate enough to offset the yield pick-up. Becoming a counterparty to long maturity FX swaps. But acting as a principal means implies capital costs: margins will be affected.
 Reinventing stock loan systems is hard to achieve with massive stock loan inventories of the kind sported by agent lenders. BNY Mellon, for example, has $3 trillion of lendable assets scattered across 31 markets. We now know it is impossible – a major agent lender tried it, and was forced to abandon the idea.
 DTCC Clarient, and Markit/Genpact.
 The debate is no longer whether CSDs and particularly CCPs are necessary. It is whether they are too powerful. Or, to put it another way, whether they are too weak to absorb the scale of the capital and liquidity risks they are now assuming. The important policy debates in regulation now are not about banking any more. They are about whether financial market infrastructures will create a collateral crunch; about whether a failed infrastructure can be recovered and, if not, resolved; and about whether financial market infrastructures should be forced to comply with a single set of global standards dictating how they govern their affairs, and provide their services, and especially on how they identify, disclose and manage their risks.
 J.P. Morgan Asset Management from BNY Mellon to J.P. Morgan; Threadneedle from J.P. Morgan to Citi; and Old Mutual from RBC to Citi.